Data & AI

Data sovereignty for data platforms and BI

Where is your data stored, who can access it, and who operates the platform? We build data platforms and BI solutions in such a way that you can answer these questions at any time: in your own data centre, in a hybrid environment, or in the cloud.

Microsoft Solutions Partner with all six designations. Data platforms and BI for organisations in Germany and Switzerland.

Data sovereignty

Data sovereignty: when "Data in the EU“ is not enough

Patient data. Client files. Design drawings. Payroll data. For data of this kind, the standard cloud is often not an option. And simply having a data centre in Frankfurt does not make it secure.

Data sovereignty means more than just a storage location. You also need to know what legal obligations your provider is subject to, who has technical access, and how you can demonstrate this to regulators, clients and customers.

These questions should be asked at the start of a data project. If you only ask them after choosing your tools, you’ll end up having to make changes later on.

Typical starting point

Does that sound familiar to you? Have you heard of it before?

Data protection has brought the cloud project to a halt

The specialist department wants reports; IT has planned a cloud platform; and then comes the veto. We’ll show you which data can still be moved to the cloud and which must remain in-house.

A client requires proof

Where is the data processed, who has access to it, and how is this documented? We are building the platform in such a way that you can answer these questions with a single document rather than by setting up a working group.

Clients or patients ask about

Law firms, medical practices and consultancy firms must not simply send confidential documents via a cloud service. Local processing and automated anonymisation help to ensure that professional confidentiality is maintained, even when data is analysed digitally.

Power BI is set up, but the cloud isn’t

Your team is familiar with Power BI, but the data must not leave the network. That’s where the Power BI Report Server comes in, or a European alternative such as TARGIT.

The BI solution is getting on a bit

Reporting has been carried out locally for years and is due to be modernised without relinquishing control over the data. We are modernising the system either in our own data centre or gradually adopting a hybrid approach.

AI, yes, but not via US services

The management wants to use AI, whilst the data protection team does not want to send any data to cloud services. Local language models make both possible.
Operating models

Cloud, hybrid or on-premises: four operating models

Cloud or on-premises? Usually both.

Sovereignty is not a simple yes-or-no question. There are several levels between the public cloud and your own data centre. Your data will determine which one is right for you.
Stage 1:

Public cloud with clear rules

Data is stored within a defined region, services are connected privately, and you manage the keys yourself. For many types of data, this is sufficient.
Stage 2:

Top-class cloud services

Major cloud providers offer options with additional controls. However, the provider and its jurisdiction remain the same. This is rarely the solution for highly sensitive data.
Stage 3:

Hybrid: a combination of cloud and on-premises

Sensitive data remains on-premises, whilst less critical data is stored in the cloud. The dividing line is determined by your data classification.
Stage 4:

Fully on-premises

The data platform, BI and AI run in your data centre, with no external connection if required. No data leaves your network.
Four questions that shape your architecture.

Data sovereignty in four dimensions

Data sovereignty is more than just a storage location. We address four questions before deciding on a platform and tools.

Where is the data stored?

In our own data centre, with a hosting provider of your choice, or in a cloud region in Germany, the EU or Switzerland. We specify where data is stored and where it is processed. These are not always the same locations.

Who can access it?

From a technical perspective, this concerns roles, permissions and proprietary keys. From a legal perspective, it concerns the question of which law a provider is subject to. Both aspects should be taken into account in the assessment.

Who runs the platform?

Your IT department, a service provider or the cloud provider. When it comes to operations, you also decide who installs updates, who has access to logs and who intervenes in an emergency.

Will you come back out?

Open formats, documented data models and interchangeable components keep your options open. This means you can still switch operating models or providers at a later date.
Building blocks

Data platform and on-premises BI: the right technology

The right tool for every stage.

TARGIT: BI from Europe

TARGIT is based in Denmark and can be run entirely in-house, in the cloud or in a hybrid environment. It is our top recommendation when analytics must remain on-premises.

In-house data platform

Data warehouses and data models based on Microsoft SQL Server and Analysis Services, hosted in your data centre. Tried and tested, well documented and with no reliance on a cloud service.

Power BI in our own data centre

With the Power BI Report Server, reports and data remain in-house. You can read about what it can do and where its limitations lie in the next section.

A confidently configured cloud

Where the cloud is the right fit, we deploy Microsoft Fabric and Azure with a specified region, private network connections and our own keys.

Anonymisation of documents

Depending on the user’s permissions, the same dashboard displays either real names or anonymised values. Contracts, legal documents and medical reports can also be automatically anonymised before they are processed further.

AI without the cloud

On the sovereign data platform, language models also run entirely locally.

TARGIT: Business Intelligence On-premises in Europe

If analytics cannot leave the premises, TARGIT is usually our first recommendation. The manufacturer is based in Denmark, and the software runs entirely in your data centre.

What TARGIT can do

Why this matters for data sovereignty

Your BI environment is not dependent on a US provider. And you decide for yourself whether TARGIT runs on-premises, in the cloud or in a hybrid environment – even at a later date.

Who TARGIT is suitable for

For organisations that want more than just standard reports: interactive analyses for specialist departments, delivered entirely in-house and independently of US manufacturers.

Power BI On-Premises using the Power BI Report Server

Does your team use Power BI, but are you not allowed to store the data in the cloud? The Power BI Report Server delivers reports from your own data centre, without the need for a gateway to the cloud.

What he can do

Where it reaches its limits

Dashboards, Q&A, apps, real-time streaming and report creation in the browser are only available in the cloud service. New features are released three times a year rather than monthly.

What it requires

The Report Server requires its own, well-equipped environment. To ensure stable operation, we recommend at least 8 CPU cores and 32 GB of additional RAM, along with Windows Server 2019 Datacentre or Windows Server 2025. Many organisations use this deployment as an opportunity to modernise both SQL Server and Windows Server at the same time.

How much it costs

From SQL Server 2025 onwards, the Report Server is already included in the Standard Edition. In earlier versions, the Enterprise Edition with Software Assurance was required for this. In Microsoft Fabric, it is only included from an F64 capacity upwards, which is usually only worthwhile if Fabric is already in use. A Power BI Pro licence is only required by those who publish reports. Those who are solely readers do not need one.

Who it’s suitable for

For organisations that, when Power BI want to stick with the status quo and, above all, need traditional reporting. Anyone who wants more analytics and self-service in their own data centre will find that TARGIT are usually better off.

Local AI

AI on your sovereign data platform

A sovereign data platform is the best foundation for AI that does not share any data externally. The language models then run in your own data centre, right alongside your data.

Making knowledge available for internal use

Assistants search through and summarise internal documents without the content leaving the network.

Analysing and anonymising documents

Contracts, legal documents or medical reports are processed locally and automatically anonymised where necessary.

No need to go via US services

No cloud account, no data transfer, no questions about the provider’s jurisdiction.
Legal framework

Requirements from Germany, the EU and Switzerland

The rules that apply depend on your data and your sector. These are the requirements we encounter most frequently in our projects.
Specification Classification
GDPR: The legal basis for all personal data within the EU, including the issue of transfers to third countries.
Swiss Data Protection Act (revDSG): applies to organisations in Switzerland and has its own rules regarding disclosure abroad.
NIS2: requires operators of important and particularly important facilities to implement risk management and provide evidence, including for their data platforms.
EU Data Act: makes it easier to switch between cloud providers and thus raises the issue of portability.
Professional duties and obligations of confidentiality: for example, in the healthcare sector, in legal consultancy or in the case of confidential assignments.

We provide technical and architectural advice. We coordinate the legal assessment with your data protection team and your legal department.

Why collana

We recommend that, what suits your data

Cloud and on-premises solutions from a single source

We build platforms on Microsoft Fabric and Azure, as well as in our own data centre. That’s why we don’t need to sell you a particular operating model; instead, we can recommend the one that best suits your requirements.

Expertise in the Microsoft environment

As a Microsoft Solutions Partner holding all six designations, we know exactly how to adjust the settings for region, network, keys and permissions. And we know where the limits lie.

A European alternative in the portfolio

With TARGIT, we have a BI solution from Denmark in our portfolio that runs entirely on-premises. So you have a genuine choice, rather than just one option from the same vendor.

A look at Germany and Switzerland

With offices in Germany, Switzerland and Poland, we are familiar with the requirements of the GDPR and the Swiss Data Protection Act from our own projects.

From strategy to the live platform

Data strategy, platform, BI and AI are all provided from a single source. You have a single point of contact rather than a chain of service providers.
Procedure

Start small, make a confident decision

Clarify requirements

During the free initial consultation, we’ll assess your data and requirements. Afterwards, you’ll know which operating models are suitable and which aren’t.

Define the target architecture

In the Data Strategy Assessment We develop the target architecture and specify which data is stored where and who has access to what.

Building a database

In the Data Sprint We will implement the platform in a few weeks’ time, including documentation for audit and operation.

Hand over and provide evidence

You will receive a documented platform, training for your team and a smooth handover to your organisation. Permissions, data flows and storage locations remain verifiable.
FAQ

Data sovereignty: common questions

What does data sovereignty mean in the context of data platforms?

Data sovereignty means that you decide where your data is stored, who can access it, who operates the platform, and how you can switch providers. A data centre in the EU only covers the first point.

What is the difference between data sovereignty and data jurisdiction?

In everyday language, the two terms are usually used interchangeably. Data ownership emphasises control over one’s own data. Data sovereignty, on the other hand, also encompasses independence from individual providers and legal jurisdictions.

Is a cloud region in Germany sufficient for sensitive data?

For many types of data, yes; for some, no. The key factors are which jurisdiction the provider falls under and whether your requirements fundamentally preclude external access. We’ll clarify this based on your data classes.

Is it possible to run business intelligence entirely on-premises?

Yes. TARGIT, Microsoft SQL Server with Analysis Services and the Power BI Report Server run entirely within our own data centre, with no external connection if required.

Is there an on-premises version of Power BI?

Yes, using the Power BI Report Server. It hosts Power BI reports in your own data centre. However, dashboards, Q&A and real-time streaming are not available there.

How much does Power BI On-Premise cost?

From SQL Server 2025 onwards, the Power BI Report Server is already included in the Standard Edition. In Microsoft Fabric, it is only included from an F64 capacity upwards. A Power BI Pro licence is only required for those who publish reports; readers alone do not need one. In addition, server resources and, where necessary, a current version of Windows Server are required.

Is there an on-premises version of Microsoft Fabric?

No. Microsoft Fabric is a cloud-only service. For operation in your own data centre, SQL Server with Analysis Services, the Power BI Report Server or TARGIT are suitable options.

Can you use AI without the cloud?

Yes. Language models can be run entirely within your own data centre. This means that documents and data remain in-house, even when using AI applications.

Is on-premises more expensive than the cloud?

Not necessarily. If you already have a data centre and SQL Server licences, the additional costs are often lower than expected. However, you will be responsible for operations and updates yourself. We have compared both options for your situation.

Do we need to replace our existing BI solution?

Not necessarily. An existing platform can often be adapted to meet your data sovereignty requirements. We will determine during the assessment whether it makes more sense to adapt the existing platform or build a new one from scratch.

Do we have to choose between the cloud and on-premises?

No. Often, some of the data remains on-premises, whilst less sensitive data is processed in the cloud. We draw the line based on your data classification.

Does this also apply to companies in Switzerland?

Yes. We work with organisations in Germany and Switzerland and comply with the relevant data protection legislation. Cloud regions in Switzerland are also an option.

Which operating model? Does this match your details?

During the free initial consultation, we’ll review your data and requirements together. You’ll find out whether your current or planned platform meets your requirements and, if necessary, what changes might need to be made.